Workflow for food classification and authenticity using yerba mate and high-resolution GC/Q-TOF

Significance of the topic

Food fraud poses serious economic and safety risks across the supply chain. Mislabeling, dilution or adulteration of premium products undermines consumer trust and quality control. Analytical workflows that combine comprehensive chemical profiling with advanced data classification are vital for ensuring authenticity and detecting contaminants in complex food matrices.

Objectives and study overview

This study demonstrates a robust workflow for classifying and authenticating commercial yerba mate brands. The goals were to develop and validate a high-resolution GC/Q-TOF–based method, build statistical models to distinguish among three brands, evaluate model performance with intentionally adulterated samples, and identify characteristic flavor compounds and environmental contaminants.

Methodology and instrumentation

Samples of three yerba mate brands (A, B, C) were extracted by a standard QuEChERS protocol. Analytical separations were performed on an Agilent 7890 GC coupled to a 7250 high-resolution Q-TOF MS operating in full-scan mode (m/z 45–650, 5 Hz acquisition). Retention indices were calibrated against an alkane ladder. Data processing steps included deconvolution with SureMass, library matching against NIST17.L, and accurate-mass confirmation via the ExactMass feature.

Data files were imported into Mass Profiler Professional (MPP) 15.1 for alignment, normalization, filtering, principal component analysis (PCA) and differential analysis (fold-change, volcano plots). Classification models (PLS-DA and SIMCA) were built in MPP, then exported to Classifier 1.1 for model visualization and unknown sample assignment.

Main results and discussion

• PCA confirmed tight clustering of replicates and clear separation among brands.
• Volcano plot and fold-change analyses highlighted multiple flavor-related lactones and furanones (e.g., 3-hydroxy-5,6-epoxy-β-ionone, α-angelica lactone) as markers distinguishing brands A and C.
• Several polycyclic aromatic hydrocarbons and environmental contaminants were also found to vary by brand.
• SIMCA models outperformed PLS-DA, achieving unambiguous classification of pure and adulterated samples, even at 5 % mixture levels.

Benefits and practical applications

• High-resolution chemical profiling enables simultaneous authentication, quality control and contaminant screening.
• Workflow supports rapid classification of unknown samples without needing full MPP processing.
• Applicable to routine QA/QC in food industry and regulatory laboratories for detecting low-level adulteration.
• Provides insights into characteristic flavor compound profiles for product differentiation.

Future trends and potential applications

• Integration of machine-learning algorithms for automated feature selection and improved model robustness.
• Extension of the workflow to other high-value foods (oils, juices, spices) and botanical matrices.
• Development of real-time or portable GC/Q-TOF systems for in-field authenticity screening.
• Building comprehensive spectral and chemometric libraries to support rapid identification and cross-validation across laboratories.

Conclusion

This study presents a streamlined, high-resolution GC/Q-TOF and chemometric workflow for authenticating yerba mate brands. The method successfully differentiated commercial samples, identified key flavor markers and contaminants, and detected low-level adulteration with high confidence. The approach offers a powerful platform for routine food authentication and quality assurance.

References

1 The Good Scents Company. Flavor and Fragrance Database. 2020.
2 PubChem. National Center for Biotechnology Information. 2020.
3 Arctander S. Perfume and Flavor Chemicals (Aroma Chemicals), Vol. 1; Lulu.com: Morrisville, NC, 2019.
4 Flament I. Coffee Flavor Chemistry; 2nd ed.; 2002.
5 Leffingwell J C, Young H J, Bernasek E. Tobacco Flavoring for Smoking Products; 1972.
6 Morata A. Red Wine Technology; 2019.